Chủ Nhật, 28 tháng 4, 2013

Remember hearing that 90% of Americans want gun control? Well, those claims were pig excrement.

Economist John Lott eviscerates the propagantician known as "Lies, Damn Lies and" Nate Silver (which is the nickname he prefers, I hear):

Despite assurances by the likes of Nate Silver that these are solid polls showing "Overwhelming majorities of 80 to 90 percent of the public say they favor background checks," 

I have previously noted my skepticism of claims.  To me, it wasn't too surprising that the Senate voted down the gun control bill about 10 days ago.  My concern is that people were really just being asked about whether they wanted to keep criminals from getting guns, not about the particular legislation being voted on by the Senate.  Well, now there is another poll by the PEW Research Center that I think is much more accurate.  It asks people whether they are happy that the Senate gun control bill was stopped.  Apparently, both Republicans and Independents are generally happy that it was stopped.  My guess is that Republicans should pay a lot more attention to what Independents and Republicans wanted than Democrats who wouldn't never have voted for the Republicans anyway.  It looks to me that Republicans voted the way that there constituents wanted.  So Republicans shouldn't really care that among all voters the poll showed support of 47 to 39 percent.  They should look at the results by political affiliation.

Many, such as the New York Times, paint a picture of Senators [like Max Baucus] who both simultaneously opposed the will of 90 percent of their voters and at the same time quake in fear of the NRA... Note in Baucus' case, he is retiring and yet he still voted against the so-called "universal background check" bill.  Might [the Times'] Mr. Nocera re-examine his piece?

By the way, the Nate Silver analysis of how Senators would vote on the Manchin-Toomey bill has sone major problem, he uses a very unreliable measure of gun ownership from Behavioral Risk Factor Surveillance System.  If you use a bad survey, you get a lot of noise and it makes it more likely that you won't get a statistically significant result.  I also like the way he runs a regression after the fact rather than predicting how people were going to vote before the Senate vote.  Letting him play around with different factors (we have no idea how many regressions that he ran before he decided to report the one that he did) makes the predictive power of that regression pretty useless.

It's back to the propaganda drawing board for you, Mr. Silver!


Hat tip: BadBlue Gun News.

Không có nhận xét nào:

Đăng nhận xét